Various iron-ore direct reduction processes have been developed but have not been commercialized because of the problem of too rapid attack of the refractory lining of the reaction vessel by the high-FeO slag. One attempt to circumvent this problem has involved water cooling of the vessel lining to provide a solidifed scull of the high-FeO slag for containment of the slag. However, when attempts were made to scale up this solidified-scull system from pilot to commerical size, it was found that water cooling requirements became excessive and produced an unacceptable energy loss. Examples of such processes are to be found in U.S. Pat. No. 3,264,096 and in the so-called INRED process developed by Boliden Kemi AB of Sweden.
Pertinent to the invention is prior recognition that certain refractory materials now in an experimental but promising state can be made inherently porous to molten iron or steel. Sintered alumina, sintered magnesia and sintered magnesia-chromia systems have been recognized in this regard with sintered alumina being presently the most promising such material for purposes of the present invention.